# 136

Dear Meervat R. Altaie, Marwa M. Dishar

I read your paper:

Applying the Analytical Hierarchy Process to Identify the Challenges and Priorities of Reconstruction Projects in Iraq

My comments

1- You really address a new and important subject as is determining the weights of uncertain and subjective weights for Irak reconstruction

2- You use the AHP method, that, in my opinion is wrong, because it was not built to deal with interconnected criteria as you have here. This was expressly mentioned by Saaty, the creator of the method, and thus, it is not a problem with the AHP method but with its correct application.

AHP uses a linear hierarchy that is not applicable to this problem where there exist transversal relationships as well as vertical ones.

Within cluster ‘Cost overrun’ you have criteria 1 and 4 that are related, since ‘Difficulty in developing a cost plant’ is related with ‘Unrealistic financial cost estimates. Obviously, if you have unrealistic financial cost, there is difficulty in developing a cost plant

And between clusters or transversal relations, there are certainly relations between diversify number of stakeholders in cluster ‘Scope’, and unrealistic finance and cost estimate in cluster ‘Cost overrun’ and dependency of importing building materials in cluster ‘Low quality’, among others.

This fact is aggravated in AHP because its procedure of selecting criteria without previously having the alternatives they must evaluate, which is against common sense.

3- If you observe the values of all criteria weights, they are quite similar, which means that there are small differences between criteria importance. Needless to say, it complicates the computation because of lack of definite advantage of one criterion over others

4- The pair-wise comparison is by far an arbitrary method without any mathematical support and lack of common sense, since it is based on intuitions and feelings.

5- It is hard to believe that criterion ‘Difficulty in identifying the needs of the affected population’, has by far the lowest value, when common sense says that it is the most important criterion, meaning that needs of affected people is secondary to repair buildings. What is more important, buildings or people? This is a clear absurd consequence of pair-wise comparison.

Same for criterion ‘Difficulty in developing a cost plan’, especially considering that the three first criteria of this cluster aim at have a reliable cost plant. Without it, nothing can be done

6- In page 5 “Check the consistency ratio, which should be at most 10%”

Why? Why there must be consistency? Because AHP says so? Where is the proof that this is a sine-qua-non condition?

7- In determining priorities why do you use the Eigen Value and the Geometric Means? One is enough

8- In page 5 “comparison, meaning that there is no contradiction in experts' opinions”

That is only because AHP software does not allow to continue until the DM modifies his estimates, and this is forced by a formula. Therefore, estimates made in good faith by the DM must be modified to satisfy a formula? Where is the rationality in this? What if the DM refuses to correct himself since no reasons are given? He cannot ‘solve’ his problem.

9- Criteria or factors as you call them, are not axes. The axes are the different alternatives or option

You can easily see this in a graphic on linear algebra

10- In page 7 “This depends mainly on the unrealistic schedule (40%), followed by the difficulty of controlling activity duration (20.6%)”

Did you perform the planning and scheduling of the whole project considering all alternatives?

That is, did you build the Critical Path Method and used the Gantt chart as well as the Earned Value Analysis, to be able to measure and control cost and delays?

In my opinion, you cannot solve this rather complex scenario using only a MCDM method, you need to plan and schedule before that. Even you need to establish relationships between alternatives order to schedule your resources like specialized manpower, concrete, artists, etc.

I hope these comments may help

Nolberto Munier